explanation condition number in GWR

1171
5
02-11-2013 12:11 PM
New Contributor II
Hallo

I´m new in statistics an I write my master thesis for the university, there I use GWR and in the GWR there is a condition number which check local collinearity. These number schould be < 30 ok!
http://resources.arcgis.com/en/help/main/10.1/index.html#//005p00000031000000

The GWR works - because of the high bandwidth the condition number is ok.

- But how dose the condition number in arcgis works?
- How is the condition number computed?

I need an explanation for my master thesis, I´m an autrian guy so unfortunatley the englisch word condition number says nothing to me.

Is it common or should it be explain it in the case of GWR?

Thanks!
Peter
Tags (2)
5 Replies
New Contributor II
Hallo

I´m new in statistics an I write my master thesis for the university, there I use GWR and in the GWR there is a condition number which check local collinearity. These number schould be < 30 ok!
http://resources.arcgis.com/en/help/...00000031000000

The GWR works and because of the high bandwidth the condition number is ok.

- How is the condition number computed in theory in arcgis?

- Is it common or should it be explain it in the case of GWR in the master thesis what do you mean?

I´m an austrian guy so unfortunatley the englisch word condition number says nothing to me.

Thanks!
Peter
New Contributor II
Is it local or global?
New Contributor II
I have found out that the ArcGIS GWR tool provides the condition number for the local X'WX matrices.

But only these two short words!

- Can anyone here specify the local X'WX matrices in more detail?
- Or give me some search intentions to find an short right explanation of the condition number with the local X'WX matrices?

Thanks!
New Contributor II
I have found out that there are future readings of these topic in Belsley DA, Kuh E and Welsch RE (1980, 2004) Regression Diagnostics: identifying influential data and sources of collinearity, Wiley
New Contributor II
I have found out that there are future readings of these topic in Belsley DA, Kuh E and Welsch RE (1980, 2004) Regression Diagnostics: identifying influential data and sources of collinearity, Wiley